TY - GEN
T1 - Auto Removal of Bright Spot from Images Captured Against Flashing Light Source
AU - Asha, C. S.
AU - Bhat, Sooraj Kumar
AU - Nayak, Deepa
AU - Bhat, Chaithra
N1 - Publisher Copyright:
© 2019 IEEE.
PY - 2019/8
Y1 - 2019/8
N2 - An image captured using a single lens camera with the sun or flashing source of light in the background frequently exhibits a bright spot in it. Usually, we observe bright spots in images caught during sunrise or sunset. The color of the bright spot varies depending on the overall background of the scene and mainly relies upon the light reflected from the lens. In several instances, bright spot, haze, or flare can arise anywhere in the image. The phenomenon occurs due to the light reflected or scattered by the imperfection of the lens and improper angle of incidence of light. One can avoid by taking the picture at a proper angle such that the spot is positioned directly on the source of light which necessitates expertise. Manual intervention is exacted for images bearing a bright spot; however, it demands the understanding of extra software tools. Moreover, manual editing becomes a time-consuming process for an extensive quantity of images. In the proposed work, we address this issue by determining the bright spot automatically and its removal by filling the gap. We experiment using available feature detection approaches for spot detection and well known in-painting methods to fill the spot area. As a promising method, we select speeded up robust features to identify the bright spot, and exemplar-based in-painting method to fill the gap in that location. We test the proposed method on more than 30 images captured against the flashing source of light. Experimental results on bright spot images reveal that the proposed method is efficient and corrected images appear natural without any artifacts.
AB - An image captured using a single lens camera with the sun or flashing source of light in the background frequently exhibits a bright spot in it. Usually, we observe bright spots in images caught during sunrise or sunset. The color of the bright spot varies depending on the overall background of the scene and mainly relies upon the light reflected from the lens. In several instances, bright spot, haze, or flare can arise anywhere in the image. The phenomenon occurs due to the light reflected or scattered by the imperfection of the lens and improper angle of incidence of light. One can avoid by taking the picture at a proper angle such that the spot is positioned directly on the source of light which necessitates expertise. Manual intervention is exacted for images bearing a bright spot; however, it demands the understanding of extra software tools. Moreover, manual editing becomes a time-consuming process for an extensive quantity of images. In the proposed work, we address this issue by determining the bright spot automatically and its removal by filling the gap. We experiment using available feature detection approaches for spot detection and well known in-painting methods to fill the spot area. As a promising method, we select speeded up robust features to identify the bright spot, and exemplar-based in-painting method to fill the gap in that location. We test the proposed method on more than 30 images captured against the flashing source of light. Experimental results on bright spot images reveal that the proposed method is efficient and corrected images appear natural without any artifacts.
UR - http://www.scopus.com/inward/record.url?scp=85081987478&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85081987478&partnerID=8YFLogxK
U2 - 10.1109/DISCOVER47552.2019.9007933
DO - 10.1109/DISCOVER47552.2019.9007933
M3 - Conference contribution
AN - SCOPUS:85081987478
T3 - 2019 IEEE International Conference on Distributed Computing, VLSI, Electrical Circuits and Robotics, DISCOVER 2019 - Proceedings
BT - 2019 IEEE International Conference on Distributed Computing, VLSI, Electrical Circuits and Robotics, DISCOVER 2019 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 3rd IEEE International Conference on Distributed Computing, VLSI, Electrical Circuits and Robotics, DISCOVER 2019
Y2 - 11 August 2019 through 12 August 2019
ER -